基于局部特征尺度分解的齿轮故障诊断方法研究

发布时间：2019-01-20 08:06

【摘要】：齿轮是机械设备中重要的连接和传动部件，在机械设备运行过程中发挥有重要的作用。但由于齿轮本身结构的特点，是特别易受损害和出现故障的零部件，如果不及时发现故障，将会给整个生产和社会造成很大的损失。可见齿轮故障诊断研究的重要性。齿轮故障诊断的关键是从齿轮的振动信号中提取故障特征，信号分析与处理是提取故障特征最常用的方法。然而，齿轮发生故障时，其振动信号大多是非平稳、非线性的时变信号，这就要选择合适的信号分析方法。本文正是针对齿轮故障振动信号的非平稳性及其多为多分量的调制信号之和等特性，将一种新的自适应时频分析方法—局部特征尺度分解方法(Local characteristic-scaledecomposition，简称LCD)引入到齿轮故障诊断中，进一步将该方法与倒频谱、能量矩、双谱等方法相结合应用于齿轮故障诊断中，并取得了较好的分析效果。本文的主要研究内容如下： 1、论文提出了一种新的自适应时频分析方法—局部特征尺度分解方法(Localcharacteristic-scale decomposition，简称LCD)，通过对其理论本身进行剖析研究和对仿真信号进行分析，指出了LCD方法有一定的优越性。同时，LCD方法也有一定的缺陷和不足，论文对其进行了改进，并且改进后的LCD方法对信号进行分解得到的分量有更好的光滑性。 2、针对齿轮故障振动信号大多数为若干的调幅调频信号之和这一特点，将基于B样条函数的局部特征尺度分解方法(B spline-based Localcharacteristic-scale decomposition，，简称BLCD)和倒频谱应用于齿轮故障诊断中，有效的提取了故障齿轮的故障特征。 3、把基于三次样条函数的局部特征尺度分解方法和能量矩相结合应用于齿轮的故障诊断，通过对正常齿轮和断齿齿轮进行分析，验证了该方法的有效性。 4、把基于有理样条函数的局部特征尺度分解方法(RLCD)和双谱应用于齿轮的故障诊断中，从双谱图中有效的提取了故障齿轮的故障特征。这也为齿轮的故障诊断提供了一种新的方法。
[Abstract]:Gear is an important connection and transmission part in mechanical equipment, which plays an important role in the operation of mechanical equipment. However, because of the characteristics of gear structure, it is especially vulnerable to damage and fault parts. If failure is not found in time, it will cause great losses to the whole production and society. This shows the importance of gear fault diagnosis. The key of gear fault diagnosis is to extract fault features from gear vibration signals. Signal analysis and processing are the most commonly used methods to extract fault features. However, when gear failure occurs, the vibration signal is mostly non-stationary and nonlinear time-varying signal, so we should choose the appropriate signal analysis method. In this paper, aiming at the nonstationarity of gear fault vibration signal and the sum of multi-component modulated signals, a new adaptive time-frequency analysis method, local eigenscale decomposition method (Local characteristic-scaledecomposition,), is proposed. LCD) is introduced into gear fault diagnosis and applied to gear fault diagnosis with the combination of cepstrum, energy moment, bispectrum and so on, and good results are obtained. The main contents of this paper are as follows: 1. A new adaptive time-frequency analysis method, called local feature scale decomposition (Localcharacteristic-scale decomposition,), is proposed in this paper. By analyzing the theory itself and the simulation signal, it is pointed out that the LCD method has some advantages. At the same time, the LCD method also has some defects and shortcomings. The paper improves it, and the improved LCD method has better smoothness to the components of signal decomposition. 2. In view of the fact that the vibration signals of gear fault are mostly the sum of amplitude modulation and frequency modulation signals, the local characteristic scale decomposition method (B spline-based Localcharacteristic-scale decomposition,) based on B-spline function is proposed. BLCD) and cepstrum are used in gear fault diagnosis, and the fault features of fault gears are extracted effectively. 3. The local characteristic scale decomposition method based on cubic spline function and the energy moment are applied to the fault diagnosis of gear. The validity of this method is verified by analyzing the normal gear and broken gear. 4. The local feature scale decomposition method based on rational spline function (RLCD) and bispectrum are applied to the fault diagnosis of gears, and the fault features of faulty gears are extracted effectively from the bispectrum. It also provides a new method for gear fault diagnosis.
【学位授予单位】：湖南大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TH165.3

【参考文献】